A large mobile network operator reduced churn by 12-18% through churn prediction model on post-paid customers
Overview
TransOrg Analytics helped a client to predict post-paid customer churn one month in advance to strategically plan and deploy retention & engagement actions
Solution
- Predictive modeling to estimate churn probability a month in advance using random forest
- Customized models for each telecom operating zone (circle) to capture local behaviors
- Review model performance (coverage, accuracy, trend, opportunity sizing) every month
Impacts
Accurately identified churners one month in advance with an accuracy/coverage of 75%
Reduced churn in key circles by 12-18%
